7. Project Summary/Abstract Peripheral nerve stimulation (PNS) in MRI results from electric fields induced by the switching of gradient coil, which may result in stimulation of the largest nerves in the body (large diameter nerves are easier to excite than small ones). The use of current generation of Gmax=80 mT/m, Smax=200 T/m/s whole-body MRI gradients is largely constrained by PNS rather than amplifier power, mechanical issues or heat removal and specialty coils such as the Gmax=300 mT/m, Smax=200 T/m/s ?MGH Connectome? coil can only be fully used within a fraction of its operational parameter space. Impacting these PNS limitations will allow faster imaging, higher resolution and reduced distortions in many sequences routinely used for research and in the clinic for head/neck as well as body imaging, such as EPI, DWI, bSSFP, RARE and PROPELLER. Head-only (HO) gradient inserts have higher thresholds but their latest generation are also PNS limited. Additionally, most neuroimaging research studies and nearly all clinical studies use whole-body (WB) gradient systems. In this program, we develop a gradient design tool with explicit PNS constraints and validate the PNS benefits by experimental tests of optimized WB and HO designs. The state-of- the-art boundary element (BEM)-stream function (SF) approach for designing the winding patterns of gradient coils optimizes the magnetic field subject to electrical, mechanical and thermal constraints, but ignores the primary limiting factor; PNS. Although design rules-of-thumb exist, PNS is not directly incorporated in the design step. Instead, PNS is assessed after construction of a coil prototype on volunteers. This is a costly and slow approach that allows only minimal PNS mitigation iteration. In this proposal, we build on our work modeling magneto-stimulation in full-body peripheral nerve models which takes into account: i) the coil wire pattern, ii) the detailed shaping of the induced electric fields by the tissue boundaries, iii) the dependence of the stimulation effect on the relative orientation between electric field and nerves, iv) the non-linear nerve dynamics and their differing properties depending on class (motor, somatosensory or autonomic) and branching distance from the CNS. Our preliminary results indicate that we can increase PNS thresholds by 2X for WB and 1.7X for HO designs. The cost is a moderate increase of the linearity error (5%) and inductance (32%, only required for WB designs). This shows that winding patterns intrinsically contain degrees-of-freedom that can support substantial PNS improvements if one has the tools to uncover them during the design phase. We therefore incorporate our PNS analysis into an industry-standard BEM-SF design optimization framework and validate our results by building and testing the best coil designs in a PNS threshold study of healthy volunteers.